YouTube to auto-label AI videos when creators skip disclosure

YouTube is shifting from purely creator-driven AI disclosures to a hybrid model where its own detection systems will automatically tag photorealistic synthetic content. Starting May 2026, if a creator omits a disclosure but internal signals flag significant photorealistic AI use, the platform will apply the label itself. Creators can contest the classification through YouTube Studio, but labels become permanent for content made with YouTube’s own Veo and Dream Screen tools or carrying C2PA metadata that marks it as fully generative.

The label itself is also moving from buried description text to a more visible position — directly under the player for long-form videos and as an overlay on Shorts. Unrealistic or animated AI content keeps its disclosure tucked in the expanded description. YouTube emphasizes the label does not affect recommendations or monetization eligibility.

The shift reflects how unreliable voluntary disclosure has been at scale, and signals a broader industry move toward provenance signals like C2PA as the ground truth for AI attribution rather than self-reporting. It also positions YouTube as an arbiter of what counts as AI-generated, with appeals routed back through its own tooling.